Integration of DInSAR Land Subsidence Observations with a Coupled Groundwater Flow and Geomechanical Modeling Approach

This contribution presents the methodology and results of modeling studies conducted as part of an international research collaboration that aims to develop an innovative approach to characterize from the hydrogeological point of view water-stressed Mediterranean basins experiencing land subsidence due to excessive groundwater extraction. By integrating remote sensing-derived land subsidence rates with an iterative implementation of numerical groundwater flow and geomechanical modeling, we developed an approach to improve the estimation of hydrogeological parameters, mainly hydraulic conductivity (K) and specific storage (Ss), and obtained a flow model that can better match historical groundwater level observations. To improve the characterization of K and Ss, hydraulic head measurements from groundwater monitoring wells and displacement observations from Differential Interferometric Synthetic Aperture Radar (DInSAR) datasets were utilized. This was achieved through a novel procedure using a 3-D groundwater flow simulator (MODFLOW) and a 3D geomechanical simulator (GEPS3D) in an iterative coupled approach, with spatial variations of Ss and K described as stationary Gaussian random fields.

This approach was demonstrated in the Alaşehir-Sarıgöl alluvial aquifer located in the Gediz River watershed (Turkiye), where groundwater withdrawal for vineyard irrigation and urban water demand have led to significant land subsidence of up to 10 cm/yr. The simulated hydraulic heads from the flow model were input into the geomechanical model, allowing the calculation of the displacement time series for the subsiding areas, which were then compared with the DInSAR data. The updated distribution of the aquifer system compressibility, as obtained by fitting the simulated to the observed subsidence trends, was iteratively used in the groundwater flow simulator to update the hydrogeological parameter values, thereby improving model performance. Calibrated groundwater flow models are expected to more accurately forecast transient groundwater storage depletion/accumulation and are, therefore, more reliable for water management decision-making. In addition, the outcome of the iterative procedure highlighted considerable heterogeneity in the parameter distribution, underscoring the importance of remote sensing-based land subsidence observations for constraining the parameters of groundwater flow models.

Acknowledgments: This study was funded by the PRIMA program under grant agreement No. 1924, project RESERVOIR. The PRIMA program is supported by the European Union.